a strange loss of stockfish in ccrl

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Dann Corbit, Harvey Williamson

Uri Blass
Posts: 10100
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

a strange loss of stockfish in ccrl

Post by Uri Blass »

What is the reason that stockfish played 79.e8Q+? and why only depth 1 for moves 78 and 79?

Is it possible that stockfish has time management bug in repeating time control because they test only incremental time control?

[pgn][Event "CCRL 40/15"]
[Site "CCRL"]
[Date "2022.06.26"]
[Round "830.4.448"]
[White "Stockfish 15 64-bit 4CPU"]
[Black "Berserk 9 64-bit 4CPU"]
[Result "0-1"]
[ECO "A20"]
[Opening "English opening"]
[PlyCount "167"]
[WhiteElo "3537"]
[BlackElo "3473"]

1. c4 {+0.00/1 0s} e5 {+0.00/1 0s} 2. a3 {+0.00/1 0s} Nf6 {+0.18/26 17s} 3. d3
{(e3) +0.12/38 90s} d5 {(Nc6) +0.28/26 14s} 4. cxd5 {(cxd5) -0.06/34 7s} Nxd5
{(Nxd5) +0.14/27 11s} 5. Nf3 {(Nf3) +0.00/35 10s} Bd6 {(Bd6) +0.43/26 10s} 6.
g3 {(g3) -0.08/34 8s} O-O {(h6) +0.31/29 40s} 7. Bg2 {(Bg2) -0.02/30 9s} h6
{(Re8) +0.23/32 44s} 8. O-O {(O-O) +0.00/37 10s} Re8 {(Re8) +0.14/30 15s} 9.
Nc3 {(Qc2) +0.00/34 10s} Nxc3 {(Nxc3) +0.19/26 7s} 10. bxc3 {(bxc3) +0.05/34
9s} Nc6 {(Nd7) +0.25/27 9s} 11. Be3 {(Be3) +0.08/30 8s} Ne7 {(Be6) +0.19/30
35s} 12. a4 {(c4) +0.00/31 26s} Nd5 {(a5) +0.19/28 10s} 13. Bd2 {(Bd2) +0.18/32
8s} Nf6 {(Nf6) +0.28/29 14s} 14. e4 {(e4) +0.04/33 14s} b6 {(b6) +0.19/29 17s}
15. Qc2 {(Qc2) +0.00/34 19s} Qd7 {(Qd7) +0.15/30 20s} 16. a5 {(a5) +0.00/37
19s} Bb7 {(Bb7) +0.27/31 11s} 17. Be3 {(Be3) +0.00/40 10s} Ng4 {(Rad8) +0.21/28
9s} 18. Bc1 {(Bc1) +0.00/33 11s} Rad8 {(Rad8) +0.09/31 10s} 19. h3 {(axb6)
+0.04/32 11s} Nf6 {(Nf6) +0.11/32 12s} 20. Rd1 {(axb6) +0.00/245 47s} Bc5
{(Bc5) +0.36/29 11s} 21. Bb2 {(Bb2) +0.00/43 10s} Qc6 {(bxa5) +0.23/34 51s} 22.
Ne1 {(axb6) +0.15/32 37s} bxa5 {(bxa5) +0.23/31 11s} 23. Rxa5 {(Ba3) +0.13/36
9s} Bb6 {(Bb6) +0.31/31 12s} 24. Raa1 {(Raa1) +0.00/34 12s} a5 {(Nd7) +0.21/34
32s} 25. Nf3 {(Nf3) +0.00/49 31s} Ra8 {(Nd7) +0.18/27 9s} 26. c4 {(c4) +0.00/41
8s} Nd7 {(Nd7) +0.48/28 6s} 27. d4 {(Ra2) +0.00/40 22s} exd4 {(exd4) +0.37/32
20s} 28. Nxd4 {(Nxd4) +0.00/45 10s} Bxd4 {(Bxd4) +0.37/32 8s} 29. Rxd4 {(Rxd4)
+0.00/46 15s} Nc5 {(Nf6) +0.02/32 10s} 30. Rd5 {(Rd5) +0.11/34 11s} a4 {(a4)
+0.08/37 16s} 31. Qc3 {(Qc3) +0.00/35 14s} f6 {(f6) +0.01/35 9s} 32. Qc2 {(Qc2)
+0.00/36 12s} Rab8 {(Rab8) +0.02/37 18s} 33. Rc1 {(Ba3) +0.00/41 20s} Ba8
{(Kh8) +0.01/32 9s} 34. Kh2 {(h4) +0.00/42 25s} Kh8 {(Kh8) -0.01/37 25s} 35. h4
{(h4) +0.00/45 20s} Kg8 {(Nb3) +0.00/39 40s} 36. f3 {(Kg1) +0.00/34 11s} Nb3
{(Nb3) +0.00/36 13s} 37. Rcd1 {(Rcd1) +0.00/40 10s} Qa6 {(Qa6) -0.01/37 11s}
38. Rb5 {(Rf5) +0.00/43 10s} Rbd8 {(Rbd8) +0.00/36 11s} 39. e5 {(e5) +0.00/44
16s} Rxd1 {(Rxd1) +0.00/33 4s} 40. Qxd1 {(Qxd1) +0.00/40 7s} f5 {(f5) -0.01/32
5s} 41. Qd3 {(f4) +0.00/46 25s} Bc6 {(Bc6) -0.01/37 24s} 42. f4 {(f4) +0.00/47
18s} Bxb5 {(Bxg2) -0.01/39 28s} 43. cxb5 {(cxb5) +0.00/41 15s} Qg6 {(Qg6)
-0.01/39 13s} 44. Qc4+ {(Bc6) +0.00/43 17s} Kh7 {(Kh7) +0.00/37 12s} 45. Bc3
{(Qxc7) +0.00/49 23s} Re7 {(Qb6) +0.01/37 26s} 46. Bh3 {(Bb4) +0.00/40 16s} Kh8
{(Kh8) -0.02/37 34s} 47. Bg2 {(Bg2) +0.00/45 25s} Kh7 {(Kh7) -0.01/39 10s} 48.
Qxa4 {(Bh3) +0.00/50 15s} Rd7 {(Rd7) +0.00/37 13s} 49. Qc4 {(Qc4) +0.00/56 14s}
Nd2 {(Nd2) +0.00/35 11s} 50. Qc6 {(Qc6) +0.00/62 16s} Rd3 {(Rd3) +0.01/35 18s}
51. Qxg6+ {(Qxg6) +0.00/64 17s} Kxg6 {(Kxg6) +0.00/36 11s} 52. Ba5 {(Ba5)
+0.00/68 17s} Nf3+ {(Nf3) +0.00/33 12s} 53. Bxf3 {(Bxf3) +0.00/72 16s} Rxf3
{(Rxf3) +0.00/33 8s} 54. Bxc7 {(Bxc7) +0.00/78 16s} h5 {(Rb3) +0.00/31 13s} 55.
b6 {(b6) +0.00/79 20s} Rf2+ {(Rb3) +0.01/33 25s} 56. Kg1 {(Kh3) +0.00/78 31s}
Rb2 {(Rb2) +0.00/30 10s} 57. Kf1 {(Kf1) +0.00/82 16s} Kf7 {(Kf7) +0.01/34 41s}
58. Ke1 {(Ke1) +0.00/84 15s} Ke7 {(Ke6) +0.00/32 26s} 59. Kd1 {(Kd1) +0.00/80
86s} Kd7 {(Kd7) +0.00/33 17s} 60. Kc1 {(Kc1) +0.00/83 12s} Re2 {(Rb3) +0.00/34
28s} 61. Kb1 {(Kd1) +0.00/80 13s} Kc8 {(Kc6) +0.00/31 7s} 62. Kc1 {(Kc1)
+0.00/77 16s} Kb7 {(Kd7) +0.00/32 14s} 63. Kd1 {(Bd8) +0.00/76 10s} Re3 {(Re4)
+0.00/32 13s} 64. Kd2 {(Kd2) +0.00/75 11s} Rxg3 {(Rxg3) +0.01/32 7s} 65. e6
{(e6) +0.00/75 11s} Rg6 {(Rg6) +0.00/32 9s} 66. e7 {(e7) +0.00/74 16s} Re6
{(Re6) +0.00/32 9s} 67. Bd8 {(Bd8) +0.00/76 42s} g6 {(Re4) +0.00/32 8s} 68. Kc3
{(Kc3) +0.00/76 13s} Kc6 {(Kc6) +0.01/32 14s} 69. Kc4 {+0.00/78 11s} Re4+
{(Re4) +0.00/31 8s} 70. Kb3 {(Kd3) +0.00/78 35s} Re6 {(Kb7) +0.00/32 11s} 71.
Kc3 {(Kc3) +0.00/72 8s} Re4 {(Re3) +0.00/33 18s} 72. Kd2 {(Kb3) +0.00/80 6s}
Re6 {(Kd7) +0.01/33 28s} 73. Kc3 {(Kd3) +0.00/73 4s} Re1 {(Re3) +0.00/33 10s}
74. Kd3 {(Kd3) +0.00/119 15s} Kd7 {(Re6) +0.01/31 14s} 75. Kd2 {(Kd2) +0.00/74
1s} Re4 {(Re4) +0.00/32 13s} 76. Kd3 {(Kd3) +0.00/245 3s} Rxf4 {(Kc6) +0.00/31
7s} 77. Kc3 {(Kc3) +0.00/33 1s} Rf3+ {(Rf3) +0.00/34 9s} 78. Kb2 {(Kc2) -0.13/1
0s} Rf4 {(Rf2) -0.01/34 39s} 79. e8=Q+ {(Kc3) +0.00/1 0s} Kxe8 {(Kxe8)
+10.11/26 5s} 80. Be7 {(Bg5) -3.74/1 0s} Kxe7 {(Rf2) +320.50/29 18s} 81. b7
{(Kc2) [%eval -#12,78] 44s} Rb4+ {(Rb4) [%eval #13,28] 12s} 82. Kc3 {(Kc2)
[%eval -#11,75] 21s} Rxb7 {(Rxb7) [%eval #11,32] 7s} 83. Kd4 {(Kd2) [%eval
-#10,81] 24s} Ke6 {(Kd6) [%eval #10,26] 18s} 84. Kc3 {(Ke3) [%eval -#9,98] 20s}
0-1
[/pgn]
Modern Times
Posts: 3517
Joined: Thu Jun 07, 2012 11:02 pm

Re: a strange loss of stockfish in ccrl

Post by Modern Times »

A known issue, it runs out of time at the end of a repeating time controls and plays almost instantly to avoid time loss. A code modification has been raised on fishtest, but it has not been prioritised.
User avatar
Guenther
Posts: 4584
Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon

Re: a strange loss of stockfish in ccrl

Post by Guenther »

Uri Blass wrote: Mon Aug 08, 2022 11:31 am What is the reason that stockfish played 79.e8Q+? and why only depth 1 for moves 78 and 79?

Is it possible that stockfish has time management bug in repeating time control because they test only incremental time control?

https://github.com/official-stockfish/S ... ssues/4000
https://rwbc-chess.de

trollwatch:
Chessqueen + chessica